Corps de l’article

Introduction

Many Canadian jurisdictions have adopted a legislative option that allows parties to expedite their arbitration process. In this scenario, either of the parties (employer or union) may apply to their respective government to pursue an expedited arbitration procedure that includes statutorily-imposed time restrictions. The labour arbitration system, originally designed to combat delays, costs, and inefficiencies in the traditional court structure (Budd and Colvin, 2008), has received increased criticism. The disadvantages of the current labour arbitration system are well-documented (Thornicroft, 2008; Winkler, 2010) where there is substantial literature addressing the increasing costs (Thornicroft, 2008) and time-consuming nature of the labour arbitration process (Ponak and Olson, 1992; Thornicroft, 1993, 1995b). The expedited arbitration process was originally proposed to alleviate the disadvantages of the traditional arbitration process. The reduced use of the expedited system, pursuant to the legislative options, provides some concern for the labour relations community. For instance, in British Columbia, there was a reduction from 218 to 162 of s. 104 complaints filed between 2015 and 2016 (British Columbia Labour Relations Board, 2017). We have taken this opportunity to ascertain what factors may be contributing to the decreased popularity of the expedited system.

Although there is a moderate amount of research addressing the delays and costs of traditional arbitration, there are comparatively fewer studies assessing whether the choice of an expedited arbitration process is truly more expedient. Furthermore, there is a dearth of literature that examines whether the choice of arbitration process impacts the decision outcome. In this study, we examine deficiencies in the labour arbitration literature to determine if expedited arbitration does, in fact, decrease delays in arbitration. Further, we also investigate whether the expedited arbitration process alters the result(s) in comparison to outcomes achieved through the traditional arbitration process. We examine these issues using dismissal cases gathered over a fifteen-year period in British Columbia and Ontario respectively.

Background: Costs and Time Delay

Two predominant issues are apparent in the current labour arbitration system: increased costs and lengthy delay in grievance resolution. Unlike the traditional court process, labour arbitration does not require the individual(s) lodging the complaint to fund legal representation. Where employees do retain legal counsel, this is generally funded by the union via employee-paid dues. However, both the union and employer incur considerable fees when they choose—as they most often do—to have legal representation. In particular, employers (Barnacle, 1991; Block and Stieber, 1987; Ponak, 1987; Wagar, 1994), frequently employ permanent counsel who are retained year-round and remunerated for preparatory work, including research and witness preparation, as well as representation at the hearing. Additional objective out-of-pocket costs include the arbitrator’s fees and fees associated with logistics such as room rental and transportation/travel costs (Thornicroft, 2008). Hidden costs include the time that the parties dedicate to case preparation, including lost time due to witness preparation. Indirectly, unresolved work disputes may impact the workplace culture; furthermore, where discontent can interfere with productive negotiations and increase the likelihood of a work stoppage, unresolved disputes may also detrimentally impact collective bargaining (Winkler, 2010).

A second primary criticism of the current labour arbitration framework is the delay associated with the system. Notably, these delays are increasing in length (Curran, 2017; Ponak and Olson, 1992). In the 1990s, Thornicroft (1993) analyzed the length of delays in cases over a twelve year period. During this time, Thornicroft found that delays totalled a cumulative additional delay of seven days per year. In a more recent study, Curran (2014) found evidence demonstrating that the average delay has increased significantly; whereas the average delay from grievance filing to resolution was 394.12 days in 1994, this same figure jumped to 448.50 days by 2004 and 730.03 days in 2012. These results show that the delays associated with resolving a grievance through arbitration nearly doubled during the twenty-year period covered by these two studies.

Delays in labour arbitration occur before and after the hearing. Before the parties even arrive at a decision to arbitrate, the multi-step grievance process itself is often time-consuming (Thornicroft, 1993, 2008). Pre-hearing delays usually comprise the largest portion of the cumulative delay (Ponak et al., 1996) and include the inability to coordinate dates with sought-after arbitrators, busy schedules of counsel/representatives and witnesses, as well as “conscious tactics” (Thornicroft, 1993, 2008) wherein management engages in purposive delay. A second pre-hearing delay in dispute resolution involves the consensual appointment of an arbitrator. Both parties research potential arbitrators and advocate for an arbitrator whose past decisions seem predictive of a favourable ruling; delays are thus incurred when the parties disagree on the potential arbitrator—as frequently happens (Bloom and Cavanagh, 1986; Nelson and Curry, 1981; Thornton and Zirkel, 1990).

Not surprisingly, research indicates that delays often extend beyond selecting arbitrators to include the receipt of post-arbitration awards (Thornicroft, 1993). Several factors contribute to the delay in an arbitrator’s award. First, as previously discussed, in-demand arbitrators often have very busy schedules which make it difficult to receive an award in a timely manner. Furthermore, arbitrators—unlike judges—do not typically have the administrative support personnel to provide independent research or assist with scheduling, compensation, and administrative matters (Kandel, 2002); this lack of support can naturally delay the process. Finally, the parties are often required to adjourn matters to schedule additional hearing days. Additional dates may be required due to preliminary motions, additional witnesses, complex legal arguments, and/or additional information discovered during the hearing process. The difficulties inherent in scheduling the initial hearing remain present and similarly delay additional hearing dates.

Grievance arbitration outcomes can also be delayed due to the evolution of the grievance arbitration process itself. Initially, disputes were resolved informally and quickly; however, the culture of arbitration has become more legal-centric insofar as a “creeping legalism” has entered into the process (Thornicroft, 2008; Zirkel and Krahmal, 2001). “Creeping legalism“ is a term that describes the increased use of lawyers—and corresponding legal procedures and tactics—in arbitral jurisprudence. This legalism has caused an overall delay in the arbitration process as a result of the transfer of time-consuming legalistic practices, common to the courtroom, to the arbitration setting which was heretofore unaccustomed to these types of tactics. Delays include additional hearings required to present prior cases, document production, and the use of technical arguments (Thornicroft, 2008), all of which increase the length and frequency of delays in arbitration hearings (Thornicroft, 1993).

A further compounding issue in the overall increase in delays during the arbitration process is that arbitrators are taking on increasingly complex cases that require significant time to hear and adjudicate (Weber v. Ontario Hydro, 1995). Following the Supreme Court of Canada’s decisions in Weber v. Ontario Hydro and Parry Sound (District) Social Services Administration Board v. O.P.S.E.U., Local 324, labour arbitrators have an expanded scope to hear complex legal issues. In Weber v. Ontario Hydro, the Court settled the jurisdictional issue between the two parties by arguing that the arbitrator had the exclusive ability to render all decisions relating to the labour dispute. In speaking for the majority, Justice McLachlin, as she was then, examined the Ontario Labour Relations Act to argue that the arbitrator had authority to adjudicate “all differences between the parties arising from the interpretation, application, administration or alleged violation of the agreement” (as para. 45). This ruling expanded the exclusive jurisdictional model into complex areas of private law and constitutional issues—cases that were formerly deemed too complex to be resolved through arbitration. A second landmark decision, Parry Sound (District) Social Services Administration Board v. O.P.S.E.U., Local 324, 2003 SCC 42, provided further clarification as to the jurisdiction of labour arbitrators. In Parry Sound v. OPSEU, the Supreme Court of Canada established that legislation such as human rights legislation and other employment-related statutes “establish a floor beneath which an employer and union cannot contract” (at para. 28). Effectively, the decision incorporated all “employment related statutes” into the collective agreements of unionized workplaces in Canada. Following these decisions, the volume and complexity of arbitrations increased. This detrimentally impacted opportunities for quick dispute resolutions in a system already under time and financial strain.

The immediate impact of the Weber and Parry Sound cases increased the volume of human rights-related cases heard by labour arbitrators. The nature of human rights cases, particularly those assessing accommodation principles, are complex; scenarios involving the duty to accommodate, in particular, require careful consideration of evolving case law and expanding human right statutes to determine the appropriate outcome. Furthermore, the experts required in accommodation cases often include medical experts, human rights experts, and accommodation experts, all of whom are busy, which can further exacerbate scheduling delays. In addition to the time-consuming nature of hearing complex evidence, the overall breadth and complexity of administrative matters such as scheduling, additional hearings, and required adjudication time all extend delays in the arbitration process.

The impact of a delayed decision on the relative success of either party at arbitration is mixed. Early research indicates that a delay in arbitration outcomes resulted in a disadvantage to the grievor who was less likely to win reinstatement (Adams, 1978). However, Adams’ (1978) findings were not replicated in subsequent studies (Barnacle, 1991; Ponak, 1987; Ponak and Olson, 1992). For instance, Ponak (1987) found that although there was an inverse relationship between the length of time from the incident to the award and the likelihood of being reinstated, the relationship was not statistically significant.

Further consequences for outstanding workplace disputes may result in workplace issues and financial liabilities for the employer. First, Williams and Taras (2000) found that re-integration of a reinstated employee may be difficult for the involved parties. Given these difficulties, some employers and unions settle on a compensation package in lieu of the employee returning to work. The longer the time between the dismissal and delay in arriving at arbitration, the larger this package is likely to be. Further, a delay in receiving the arbitration award may result in significant back-pay (Thornicroft, 1993).

Procedural justice focuses on the arbitration process’s system evaluation and fairness (Folger, 1977). Early research on procedural justice found that litigants’ assessment of the court procedures’ fairness was correlated with their reported outcome satisfaction and subsequent authoritative evaluations (Casper, Tyler and Fisher, 1988; Melton and Lind, 1982; Tyler, 1984). Further, Thibault and Walker (1975) found that the perception of retaining control, a key element of procedural fairness, was important to individuals. Tyler (2007) defines four elements of procedural justice: 1- participation (voice); 2- neutrality; 3- respect; and 4- trust. If grievors perceive that they did not have an opportunity to address all four elements, they were more likely to believe that procedural justice was negatively impacted when comparing expedited and traditional arbitration processes. For example, if there was a considerable difference in the number of days of hearing or the time allotted for participation, grievors may allege that they were not afforded the same opportunity to obtain procedural justice as they would have in traditional arbitration.

Expedited Arbitration: Legislation

There are a variety of processes that the parties may engage to quicken the arbitration process. First, there are alternative expedited arbitration processes where the parties agree, through the collective agreement, to expedite processes that are conducive to a quicker resolution of the workplace dispute. Furthermore, the parties may elect to introduce processes, on an individual grievance basis, that expedite the traditional labour arbitration process (Kauffman, 1992). Second, in response to the well-founded delay concerns, many Canadian jurisdictions have adopted an opportunity for the parties to apply for expedited arbitration, pursuant to the applicable legislation, to accelerate the dispute resolution process.

These statutes share the goal of expediting arbitration and adopting processes that assist in this endeavour. There are a number of differences between expedited and traditional arbitration that contribute to a quicker grievance outcome. The most obvious difference is that the expedited process is designed for a quicker resolution of a workplace dispute. As such, the legislation dictates the initial date of hearing. In Ontario, Section 49 (7) of the Labour Relations Act requires the arbitrator to hear the matter within twenty-one days of their appointment. In British Columbia, Section 104 (4) (b) of the Labour Relations Code requires the arbitrator to commence the hearing within twenty-eight days (Labour Relations Code, 1995). These requirements contrast with the traditional labour arbitration process in which there are no time restrictions imposed on the parties (unless the parties voluntarily choose to impose restrictions). These factors support the timely hearing of the matter more quickly in the expedited process than in the traditional arbitration process.

The legislation also dictates time recommendations in releasing the decision. The British Columbia Labour Relations Code requires that the arbitrator issue a decision within twenty-one days of the conclusion of the hearing. However, if the parties request, and the arbitrator deems it possible, the arbitrator may release an oral decision one day after the conclusion of the hearing (Labour Relations Code, s. 104 (7)). The Ontario Labour Relations Act requires that the arbitrator give an oral decision forthwith or as soon as possible (s. 49 (8)). Given that there are no legislative suggestions that an arbitrator release the decision within a given time frame in traditional arbitrations, these requirements are unique to the expedited arbitration process. Therefore, decisions that are heard pursuant to the expedited arbitration process will likely be released earlier compared to the traditional process. As a consequence of these legislatively-based time restrictions, there should be minimal delay between dismissal and the release of the decision compared to traditional arbitration.

A third distinctive feature of expedited arbitration is that the legislation dictates the appointment of the arbitrator. In a traditional arbitration process, the parties jointly select an arbitrator. Conversely, the respective minister appoints an arbitrator from an “arbitrator’s list” in expedited arbitration. Unlike the traditional arbitration process, the legislation does not allow for direct input from the parties, preventing the appointment of a jointly-agreed upon arbitrator.

There is a dearth of research examining the expedited arbitration process. Rose (1986) found that statutory systems are more expeditious than conventional arbitration and that the expedited process can produce financial savings for both parties. Sandver, Blaine, and Woyar (1981) explored private arbitration processes in the postal, railway, and paper industry and found that expedited systems do have decreased time delays and reduced arbitration costs. These systems addressed non-complex issues and sometimes excluded discharge cases. However, these studies did not exclusively focus on statutorily-based expedited arbitration. Given that significant time has elapsed since this publication, we want to further explore these issues within the current social, economic, and political environment.

Methods

We analyzed documents in accordance with pre-determined categories (Bryman, 2008) consistent with the methodology of similar studies including Ponak et al. (1996) and Thornicroft (1993, 1994). Specifically, we limited our analysis to discharge decisions. As per Block and Stiber (1987), we restricted our analysis to discharge decisions chiefly because: 1- the language of disciplinary matters minimizes discrepancies in contract language; 2- given that the cases often address unsatisfactory behaviour, these cases have similar issues; and 3- the outcomes can be quantified into discrete categories. When one or more grievors were included in a single decision, each grievor was coded as a separate case (Grant, 2008).

Data Collection

We analyzed arbitration awards in discharge cases decided exclusively in British Columbia and Ontario. We chose British Columbia and Ontario because these provinces offered the largest number of expedited cases to study. Examining two distinct jurisdictions also allowed us to examine an overall greater number of cases compared to focusing on one jurisdiction alone. We collected cases with decisions issued in 1997 through 2011 from the LexisNexis® database, including all reported expedited arbitration cases during this same time period. To compare an equivalent number of expedited and non-expedited cases, we used a random sample generator to select a similar number of cases. We also adopted random selection to minimize selection bias and sample selection bias (Marczyk, DeMatteo, and Festinger, 2005).

Given the parameters of this study, we did not code all cases from the traditional sample. Instead, we selected all expedited cases and comparator cases from 1998, 2000, 2005, and 2010, to allow for an even distribution over the collection period. Notably, we did code the entire expedited sample due to differences in the number of expedited and traditional cases. We used a random number generator to select the traditional cases; for instance, if there were thirty-two traditional cases involving dismissal for violence in the workplace and only ten expedited cases, we selected all ten expedited cases and used a random generator to select ten cases from the traditional sample. We included each of these randomly selected cases in the coded sample.

Finally, we conducted inter-rater reliability on the primary variables. We examined variables including: 1- whether the case was expedited or traditional; and 2- whether the grievance was granted. The results of the inter-rater reliability revealed that there was 100% agreement between the coders for these variables. The total sample size was 554.

Dependent Variables

We analyzed three dependent variables addressing delays in arbitration. These included: 1- the duration from the date of the termination to the first day of the hearing with an arbitrator (duration to hearing); 2- the duration from the last day of the hearing to the arbitrator’s award (duration to award); and 3- the total duration from the date of the discharge to the release of the decision (total duration). We measured the total duration as the sum of the duration from the date of termination to the discharge of the award, as expressed in days.

We examined two additional dependent variables including the number of days of the hearing and the length of suspension that was awarded, if applicable. The length of suspension refers to the number of days that a grievor was suspended following discharge by the arbitrator. The final dependent variable was the case outcome; specifically, whether the grievance was granted (0) or denied (1).[1] We also conducted an ordered analysis with the following categories: reinstate the grievor (0); provide a suspension with thirty or fewer days (1); provide a suspension of 31 to 120 days (2); provide a suspension with 121 days or more (3); or the termination was upheld (4). If the case indicated a suspension in weeks, we multiplied this figure by seven to arrive at the total number of days. If the suspension was provided as a number of months, we multiplied this number by thirty to arrive at the total number of days.

Independent Variables

The primary independent variable that we examined was whether the case was expedited or traditional. We categorized cases as expedited when the parties utilized Section 49 of the Ontario Labour Relations Act or Section 104 of the British Columbia Labour Relations Code. We categorized cases as traditional when they followed the provisions of their applicable collective agreement.

Control Variables

We included a series of control variables including: 1- the jurisdiction of the case (British Columbia or Ontario); 2- year of the case; 3- gender of the arbitrator; 4- gender of the grievor; 5- gender of the employer representative; 6- gender of the union representative; 7- the use of legal counsel by the union and employer; 8- presence of a policy on the matter; 9- record of the grievor; 10- industry of the employer; and 11- category of offence. We selected these control variables because research suggests that these are the factors associated with the case outcome.

Some variables had more predictive impact on arbitration outcome than others. There is mixed research regarding whether a grievor’s gender is associated with case outcome (Bemmels, 1988, 1991; Knight and Latreille, 2001; Mesch, 1995; Thornicroft, 1995a). For instance, Bemmels (1988, 1991) found that male arbitrators treated female grievors more favourably than male grievors. However, other studies demonstrate no statistical difference in how female arbitrators treated male and female grievors (Bemmels 1988, 1991). The use of legal counsel is a stronger predictor of case outcome (Barnacle, 1991; Mark, 2000; Ponak, 1987). For instance, Harcourt (2000) found that grievors were more likely to have discipline overturned when the union used legal representation and the management did not employ counsel. Notably, in the reverse situation—wherein management employed representation while the union did not—management was not more likely to succeed in the grievance. There was no statistical difference in outcome when both parties used legal counsel. The past record of the grievor may also impact the arbitration outcome (Simpson and Martocchio, 1997). Notably, a laboratory study found that where arbitrators made fact-based decisions, they were more likely to uphold discipline and/or discharge when the grievor had a past record of discipline (Simpson and Martocchio, 1997).

We coded the control variables according to the following rules. First, jurisdiction was coded as Ontario (0) or British Columbia (1). Second, the year of the case was coded dichotomously as 1997 to 2003 (0) and 2004 to 2011 (1). Third, the gender of the arbitrator, grievor, and representative were coded as male (0) and female (1). When there were multiple counsels/representatives, we coded the gender of the most senior representative. Fourth, if there was an arbitration board, we restricted our coding to the arbitrator. Fifth, we coded the use of legal counsel to include whether the employer did not use legal counsel (0) or did use legal counsel (1). We similarly coded the union’s use of legal counsel as did not use legal counsel (0) or did use legal counsel (1). Sixth, we determined that membership in a provincial law society was the determining factor in establishing whether the representative(s) had legal training. As Canadian lawyers must be members of a provincial law society to practise law in Canada, the Canada Law List provides an annual list of lawyers with active memberships. Given that the list is only released annually, we recognized that current lists may not include lawyers who were practising at the time of the case, but are no longer currently practising. This was a particular concern given the case collection period of fifteen years. To rectify this potential for errors, we consulted previous publications by the Canadian Law List to determine if the representative was a lawyer at the time of the hearing to avoid mistakenly coding lawyers as non-lawyers (Carswell, 1997). Seventh, we coded employers who did not have a policy related to the grievance as (0) and the presence of a policy as (1). Eighth, we controlled for the past record of the grievor to include a clean (0) or existing record (1). If the previous record was expunged, the grievor was considered to have a clean record. Ninth, we coded the industry of the employer according to the following sector: manufacturing (1); services (2); government (3) ; mining, forestry, and mill (4); healthcare (5); food and drink (6); transportation (7); and any other sectors (8). The tenth and final variable that we coded was the subject matter of the offence according to whether it focused on dishonesty/theft (1); assault (2); alcohol/drug use (3); insubordination (4); work performance (5); harassment (6); off-duty conduct (7); attendance (8); and other offence(s) (9).

Results

We report the descriptive statistics for the dependent and independent study variables in Tables 1 and 2, respectively. We used independent samples t-tests to test whether arbitration length differed between traditional and expedited arbitrations. Raw dependent variables were not normally distributed, such that skewness and kurtosis values varied significantly from zero; thus log-transformed variables were used in t-tests to meet assumptions of normality. Levene’s test of homogeneity of variances was non-significant for all dependent variables (p > .404), except for delay to hearing (F (1) = 7.39, p = .007); thus, we reported adjusted statistics to compare hearing delays as equal variances cannot be assumed. The sample size, as presented in Table 1, did not amount to 554 cases in all categories due to missing values in the arbitral jurisprudence, including variables addressing duration to hearing, duration to award, and the number of hearing days.

Table 1

Descriptive Statistics for Dependent Variables

Descriptive Statistics for Dependent Variables

Note: The data were split between expedited (Exp) and traditional (Trad) cases.

-> Voir la liste des tableaux

Table 2

Descriptive Statistics for Independent Variables

Descriptive Statistics for Independent Variables

-> Voir la liste des tableaux

As expected, expedited cases had significantly shorter total delays on average (M = 193.34, SD = 156.90) compared to traditional cases (M = 381.58, SD = 299.54), where t (458) = 9.34, p < .001). More specifically, expedited cases had significantly shorter delays to hearing on average (M = 108.74, SD = 95.47) compared to traditional cases (M = 257.36, SD = 225.37), where t (467) = 9.25, p < .001). Similarly, expedited cases had significantly shorter delays to award (M = 86.74, SD = 123.80) compared to traditional cases (M = 130.58, SD = 180.57), where t (476) = 3.15, p < .001). Expedited cases also had fewer hearing days (M = 3.33, SD = 3.09) compared to traditional cases (M = 3.92, SD = 4.05), although this comparison failed to reach significance, (t (507) = 1.87, p = .062), see Table 2.

Table 3

Frequencies of Arbitration Outcomes for Traditional and Expedited Cases

Frequencies of Arbitration Outcomes for Traditional and Expedited Cases

-> Voir la liste des tableaux

We used a chi-square test of association to test whether the use of a particular arbitration method was associated with different award outcomes. We present frequencies of each possible outcome for both traditional and expedited cases in Table 3. There was no significant association between arbitration method and outcome, χ2 (4) = 3.52, p = .474. Thus, although expedited cases were associated with fewer delays, there were no significant differences between expedited and traditional cases in terms of suspension length or outcome. These findings are notable as they contrast with earlier research that found that delays in grievance outcome negatively reduced the odds of reinstatement (Nelson and Udan, 1998). These differences may be influenced by an increasing awareness of this trend amongst arbitrators, the use of recognized legal tests, and appropriate outcomes for dismissal cases. Thus, it is noteworthy that the raw scores of reinstatement in Table 3 reveal that fewer grievors were awarded reinstatement in the expedited process. Rather than demonstrating a direct relationship between the type of arbitration process that the parties selected, these data may reflect that parties are more reluctant to advance to the expedited arbitration process given that arbitration cases have become increasingly complex.

Multiple Regressions

We used ordinary least squares (OLS) multiple regressions to investigate potential predictors of delays in the arbitration process and suspension length. We present the results for each of the five OLS regressions in Tables 4-8.

Table 4

Predictors of Duration to Hearing

Predictors of Duration to Hearing

Note. Predictors of delay 1, number of days until hearing. R2 adjusted for number of predictors. Manufacturing and attendance were excluded due to collinearity.

* p < .05; ** p < .01; *** p < .001

-> Voir la liste des tableaux

All final OLS models were statistically significant (p < .001), with the exception of the model predicting suspension length (p = .532). Although the R squared values were modest, at .15 to .17, this may be due to the systemic nature of delays in expedited arbitration. The systemic nature of the delays in arbitration may lead to difficulty in explaining the delay by the independent variables that this study investigated. As shown in Table 4, longer delays to hearing were associated with traditional arbitration, female gender of the grievor, and the government employment sector. The B score, -.36, suggests that fewer delays in reaching the first day of hearing, arbitration award, and total delay, may be moderately predicted by using an expedited method over a traditional method in comparison to the other predictors in the model. The B values of the greivor’s gender and the government employment sector suggest only a minimal relationship.

By comparison, greater time duration to award (Table 5) was associated with the jurisdiction of Ontario, use of counsel by both the employer and the union, the government employment sector, and offences related to theft/dishonesty. The B score, -.04 in Table 5, suggests that fewer days of delay in reaching the arbitration award can be minimally predicted by using an expedited method over a traditional method in comparison to the other predictors in the model. The B value, -.37, indicates that the jurisdiction of Ontario is moderately associated with a greater duration in receiving the arbitration award. This may reflect a greater workplace population, busier arbitrators, or a culture more accepting of delay. The use of counsel, by both the union and the employers, is minimally associated as a predictor of delay. This is consistent with earlier literature that found that the use of legal counsel does increase delays in arbitration hearings (Thornicroft, 1993). Although statistically significant, the industry categories, such as government and transportation, provided minimal prediction of the greater duration in the number of days to receipt of the arbitration award. Similarly, the relationship between the offence variables, including dishonesty/theft and work performance, and the duration of the delay were small.

Table 5

Predictors of Duration to Award

Predictors of Duration to Award

Note. Predictors of delay 2, number of days from hearing until decision. R2 adjusted for number of predictors. Manufacturing and attendance were excluded due to collinearity.

* p < .05; ** p < .01; *** p < .001

-> Voir la liste des tableaux

As evident in Table 6, longer total duration of the arbitration process was associated with traditional arbitration, jurisdiction in Ontario, the union’s use of legal counsel, and the government employment sector. The B values of the arbitration method, -.28, suggest that fewer delays in completing the arbitration process can be moderately predicted by using an expedited method over a traditional method, in comparison to the other predictors in the model. The B values in the jurisdiction, use of counsel, and government employment sector suggest a small relationship between the variables and the total duration of the arbitration process.

Table 6

Predictors of Total Duration of the Arbitration Process

Predictors of Total Duration of the Arbitration Process

Note. Predictors of total delay, number of days until decision is released. R2 adjusted for number of predictors. Manufacturing and attendance offence were excluded due to collinearity.

* p < .05; ** p < .01; *** p < .001

-> Voir la liste des tableaux

In the model addressing predictors in the numbers of suspension days (Table 7), only Ontario jurisdiction predicted longer suspensions; however, as previously discussed, this model was not significant and accounted for less than one percent of the variance in suspension length. This suggests that other factors besides the ones we examined in this study may influence suspensions in arbitration. Finally, Ontario jurisdiction predicted an increase in the total number of hearing days, as did the union’s use of legal counsel, the government employment sector, and employee offences related to theft/dishonesty, insubordination, and work performance (Table 8). The B values of the union’s use of counsel and employee offences suggest a small predictive value between these variables and the number of hearing days.

Table 7

Predictors of Numbers of Days of Suspension

Predictors of Numbers of Days of Suspension

Note. Predictors of suspension length. R2 adjusted for number of predictors. Manufacturing sector and attendance were excluded due to collinearity.

* p < .05; ** p < .01; *** p < .001

-> Voir la liste des tableaux

Table 8

Predictors of Numbers of Days of Hearing

Predictors of Numbers of Days of Hearing

Note. Predictors of days of hearing. R2 adjusted for number of predictors. Manufacturing and attendance were excluded due to collinearity.

* p < .05; ** p < .01; *** p < .001

-> Voir la liste des tableaux

Probit Regressions

We used two probit regressions to examine potential predictors of arbitration outcomes. Binary probit models are appropriate for dichotomous dependent variables since ordinal probit models are a regression method where the dependent value can be categorical. In this case, we divided the arbitration outcomes into progressive outcomes. We present the results for the binary probit regression in Table 9. The model was significant, with a pseudo R2 value of .21. Greater likelihood of a grievance denial was associated with the year of the case, female gender of the arbitrator, the employer’s use of legal counsel, union’s non-use of legal counsel, and the employee’s record. The B value for the year of the case suggests a strong predictive relationship between upholding the termination and the earlier arbitration decisions (1997-2003). That this trend weakens over time may be related to developments in jurisprudence that influence case outcomes. Interestingly, the B value, .53, suggests a strong predictive value in that the employer’s use of legal counsel was positively related to the grievance being upheld. In contrast, the union’s use of counsel was moderately related to the grievance being granted. The past record of the grievor had a small relationship with the outcome where a past record was positively associated with the success of the employer.

Table 9

Predictors of Arbitration Outcome (Binary Probit)

Predictors of Arbitration Outcome (Binary Probit)

Note. Predictors of arbitration outcome (grievance granted = 0; grievance denied = 1). “R2” here refers to Naglekerke R2 values. Chi-square value represents the likelihood ratio chi-square testing the significance of the final model against the intercept-only model. Manufacturing sector and attendance were dropped due to previous collinearity analysis.

* p < .05; ** p < .01; *** p < .001

-> Voir la liste des tableaux

We present the results for the ordinal probit regression in Table 10. The model was significant, with a pseudo R2 value of .21. Employers were more successful in earlier cases, when the arbitrator was female, when employer used legal counsel, and in the transportation sector. The B score -.43, for year of cases, suggests that earlier cases provide a moderate-to- strong predictor of the employer’s likely success. The B values of the employer using counsel, .62, and the female gender of the arbitrator, .45, also suggest a strong predictive value in comparison with other variables in the study. By comparison, the union was more successful when it used legal counsel, when the issue included “other offences”, and in the healthcare employment sector. The relationship of union counsel was moderate, where the B value was -.31, which does suggest that the use of counsel assists the union in successfully advancing their case.

Table 10

Predictors of Arbritation Outcome (Ordinal Probit)

Predictors of Arbritation Outcome (Ordinal Probit)

Note. Predictors of arbitration outcome (grievor reinstated = 0; suspension of ≤ 30 days = 1; suspension of 31-120 days = 2; suspension of ≥ 121 days = 3; termination upheld = 4). “R2” here refers to Naglekerke R2 values. Chi-square value represents the likelihood ratio chi-square testing the significance of the final model against the intercept-only model. Manufacturing and attendance offence were dropped due to previous collinearity analysis.

* p < .050; ** p < .010; *** p < .001

-> Voir la liste des tableaux

Discussion and Policy Implications

The motivation for this study was to determine whether the choice of arbitration process impacts total arbitration delays and outcomes. Arguably, the most meaningful contribution that this study offers is the finding that there is no statistically significant difference in the arbitration outcomes of expedited cases compared to traditional cases. Specifically, our research reveals that there are no statistically significant differences found in the relationship between the arbitral outcomes (when considering dismissal cases) and the chosen arbitration process. That is, expedited versus traditional arbitration did not predict whether the grievance was granted or denied. These results should lend some comfort to those that may believe that arbitration results may differ in the expedited arbitration process compared to the traditional process.

However, the expedited arbitration cases were characterized by fewer delays than traditional arbitration cases. Specifically, there were significantly fewer days of delay in obtaining a first day of hearing, in the release of the decision, and in the total delay of the overall arbitration in the expedited arbitration method. These findings are consistent with earlier research that found that delays were less, at all stages of the process, in the expedited process than the traditional process (Rose, 1986). However, other factors may influence this relationship, including the use of tripartite boards which may be less expedient (Ponak and Olsen, 1992). The relative experience of the arbitrator may also impact delays. That is to say, less experienced arbitrators typically hear expedited cases because they are typically less busy than experienced arbitrators; consequently, experienced arbitrators are more likely to hear traditional arbitration cases. Arguably, the arbitrator’s training, experience, and knowledge may impact their procedures and findings. Furthermore, since arbitrators hearing expedited cases are likely to be less experienced and have more recently completed their education, they may have been better trained in the importance of addressing delays within the arbitration process.

Overall, the expedited arbitration process was not significantly associated with fewer days of hearing than traditional arbitration. This indicates that although the expedited arbitration process may be quicker, there is little difference in the number of total hearing days. This finding is particularly important given grievors’ concerns regarding procedural justice and opportunities for voice in the arbitration process.

Our research indicated that other aspects of labour law, as they pertain to arbitration, deserve further study. For instance, there are mixed reports related to the role of gender in the outcome of arbitration (Bemmels, 1988, 1991; Thornicroft, 1995a). Further studies should address the impact of the gender of the grievor, arbitrator, and representative; this is particularly pertinent given the growing number of female lawyers and arbitrators. Further, the impact of using legal counsel should be re-examined. Our study found a positive relationship between the use of counsel, on both sides, and increased delays in the arbitration process. Further, we found a relationship between the use of counsel and the arbitration outcome. Investigating the specific impact of lawyers on the expedited arbitration process may provide meaningful contributions to this question.

Another factor that may impact delays in arbitration is the type of case advanced by the parties. Intuitively, simpler cases are likely to have quicker settlements than complex and protracted cases. Therefore, the duration of labour arbitrations may be impacted by the complexity of the case in question. Furthermore, given the reluctance of many grievors to choose an expedited process because it is does not allow for the jointly-agreed upon appointment of an arbitrator, it is possible that the expedited process itself encourages settlement in contrast to a traditional arbitration. Specifically, parties may be reluctant to leave the termination’s outcome to an arbitrator that has minimal past experience with hearings or the parties. Considering the protracted nature of labour arbitrations, these factors are worthy of further investigation.

Our results present important implications for unions, employers, and government/policy makers. First, there are important repercussions for unions and employers related to the finding that the outcomes of dismissal decisions are not statistically different based on the choice of expedited versus traditional arbitration process. Although unions and employers frequently have concerns about the potential disparity in the award result, we found that fears that the choice of arbitration process will impact these decisions are unwarranted. Furthermore, we also found that the choice of arbitrator, whether appointed by the respective provincial ministry or selected by the parties, did not result in a statistically different impact on the results in dismissal cases. This finding may be especially positive news for unions who are apprehensive that the quality of the decision may differ based upon the expedited process.

Expedited arbitration cases typically represent less of a delay than traditional arbitration cases. Therefore, adopting the expedited process should result in decreased individual and workplace disruption. As well, the timely resolution of grievances may positively impact individual grievors. Specifically, if arbitral awards are released in a quicker fashion, grievors will be able to respond appropriately to these findings. This is important in case the grievor does not return to the workplace and must move permanently to alternative employment. Reduced delays also mean that employers are likely to incur a costly reinstatement award. Finally, workers are less likely to be distracted by an on-going and unresolved labour-management conflict, which improves the overall workplace culture.

A further implication of our findings for employers and unions is that adopting the expedited process means that parties remain equally as likely to receive adequate time to present their respective cases. Although legislation does not provide time restrictions for hearing the case, our results confirm that, statistically, both parties generally do receive equal time. Therefore, our results should discount fears that procedural justice may be infringed at the expense of expediency.

Notably, our study includes a number of limitations. These include the characteristics of the individual cases, which may lead one, or both of the parties to elect the expedited process. For instance, cases that are less complex may be deemed more suitable for an expedited process. Therefore, expedited cases may be less complex in contrast to traditionally arbitrated cases. However, we sought to address this discrepancy by limiting our sample to termination cases in which legal tests addressing the arbitrator’s decision were consistent across the sample cases. Dismissal cases are often more legalistic, including the use of lawyers, which we controlled for in the study. However, the narrow scope of dismissal cases naturally limits the generalizability of the findings. That is, although these results offer guidance to the expedition of arbitrated dismissal cases, our findings are less informative for other issues related to the contract.

Another limitation is that Ontario legislation, in contrast with British Columbia, includes a narrow provision allowing the arbitrator to release the decision orally in some circumstances. The divergence in potential methods may influence the case sample within and across the two jurisdictions. Furthermore, the arbitrators appointed in expedited arbitration cases are typically less experienced than those in traditional arbitration, which may have a difficult-to-quantify impact on the outcome of the arbitration. However, we found that this possibility is reduced given that all arbitrators are influenced by the same applicable case law and judicial review process.

Conclusion

Overall, our study provides important implications for management and unions. Specifically, our results demonstrate that the expedited process is associated with fewer delays than the traditional process. Notably, the more expedient process is not associated with statistically significant differences in outcomes in comparison with the traditional arbitration process. Our findings support commentary by researchers or mediators, including Winkler (2010) among others, who advocate for alternative methods to encourage expedited arbitrations. Given the increased pressure on the traditional arbitration system, our findings suggest that there are significant advantages to expedited arbitration. Although many members of the legal community specializing in labour relations have concerns about expedited arbitration, our findings suggest that parties may want to increase their use of expedited arbitration.